Talks - Orbitronics 2022 - Mathias KLÄUI, University of Mainz

Summary

The speaker discusses experimental results in the field of spintronics, focusing on the challenges faced by experimentalists in advancing beyond current technology. They delve into the potential of spin-orbit torques and the role of orbital angular momentum, highlighting recent discoveries that suggest a significant increase in efficiency compared to traditional spin transfer torques. The talk emphasizes the importance of collaboration between theorists and experimentalists to understand the origins of observed effects and the potential for new materials to enhance performance in memory and logic devices.

Takeaways

• 😀 The speaker expresses excitement about seeing attendees in person after a period of virtual meetings.
• 🔬 The presentation focuses on experimental results and the practical challenges faced by experimentalists in the field of spintronics.
• 🌟 The speaker is from Mines and has a co-affiliation at NTNU, emphasizing the importance of understanding complex theories for practical applications.
• 🛠 There is a discussion on the limitations of spin transfer torque devices in the market, highlighting the efficiency limit of one Bohr magneton per electron.
• 📈 The potential of spin-orbit torques (SOTs) to overcome these limitations by allowing the transfer of more than one Bohr magneton per electron is introduced.
• 🔍 The experimental challenge of distinguishing between different types of torques (damping-like and field-like) is highlighted, as only the effective field's direction and strength can be measured.
• 🌀 The concept of orbital angular momentum and its potential to enhance torque efficiency beyond spin angular momentum alone is explored.
• 💡 The importance of collaboration between theoreticians and experimentalists to interpret experimental results and understand the origins of different torques is emphasized.
• 📚 The speaker references previous work and theoretical calculations that motivate the experimental approach to identifying and quantifying orbital effects in materials.
• 🛑 The experimental results show a significant increase in torques when certain materials like copper oxide are used, suggesting a strong orbital current generation.
• 🔧 The use of different ferromagnets in experiments reveals a unique dependence on the material, suggesting that the conversion of orbital currents to spin currents is material-specific.

Q & A

What is the main focus of the speaker's presentation?

The speaker's presentation focuses on experimental results related to spin and orbital angular momentum in materials, particularly the effects of spin-orbit torques and the potential for increasing efficiency beyond current limitations.

What is the significance of seeing 80 people 'alive in 3D' compared to 'tiles of Zoom'?

The speaker is expressing the value of in-person interactions over virtual meetings, highlighting the importance of live attendance for a more engaging and dynamic experience.

Why is the efficiency of spin transfer torque limited?

The efficiency of spin transfer torque is limited because it is fundamentally constrained by the efficiency of one Bohr magneton (h bar) per electron for spin transfer torque, which necessitates a certain amount of current to reverse the magnetization.

What is the potential advantage of orbital angular momentum over spin angular momentum in this context?

Orbital angular momentum offers the potential to transfer more than one Bohr magneton per electron, which can lead to increased efficiency in torques compared to traditional spin transfer torques.

What are the two fundamental mechanisms the speaker refers to in relation to spin-orbit torques?

The two fundamental mechanisms are the spin Hall effect, where spin-polarized electrons are scattered into a ferromagnet, and the inverse spin galvanic effect or Edelstein effect, where a non-equilibrium spin density forms and exerts a torque by exchange interaction on the magnetization.

Why is it challenging for experimentalists to identify the origin of a torque?

It is challenging because the only measurable quantities are the direction and strength of the effective field, which can have the symmetry of either a damping-like or field-like torque. The actual origin of the torque requires indirect interpretation and input from theoreticians.

What is the significance of the 16-fold increase in torques when copper oxide is added on top of platinum?

The 16-fold increase indicates a highly efficient generation of orbital currents that are converted into spin currents in the platinum layer, which then generate a significant spin-orbit torque acting on the underlying ferromagnet.

How does the speaker's research group address the experimental challenge of identifying orbital effects?

The group uses a combination of torque measurements and magneto-resistance effects, varying the thickness of different layers in their samples, and comparing results across multiple materials to infer the origins of the observed effects.

What is the importance of being able to manipulate iron cobalt boron instead of nickel in MRAM devices?

The ability to manipulate iron cobalt boron is important because it is preferred by companies for MRAM devices due to its compatibility with MgO tunnel junctions, which nickel does not offer.

What is the role of theoreticians in this field of research?

Theoreticians play a crucial role in interpreting experimental results, providing calculations and models that help experimentalists understand the underlying physics of the observed phenomena, and guiding the design of new experiments.

Outlines

00:00

🧲 Spin Transfer and Spin-Orbit Torques

The speaker begins by expressing excitement about the live audience and then dives into the topic of spin transfer torque devices available in the market. These devices operate by polarizing electrons and using their spin to change the magnetization of a free layer. However, they are limited by an efficiency of one Bohr magneton (h bar) per electron. The speaker introduces the concept of spin-orbit torques, which have shown potential for higher efficiency, and discusses the fundamental mechanisms behind them, including the spin Hall effect and the inverse spin galvanic effect. The talk also touches on the experimental challenges faced by researchers in distinguishing between different types of torques and the importance of collaboration with theoreticians to interpret experimental results.

05:01

🔬 Exploring the Origins of Spin-Orbit Torques

The speaker discusses the complexities of identifying the origins of different types of torques in experiments. Initially, it was believed that the spin Hall effect and the Edelstein effect produced distinct types of torques, but it has been established that both mechanisms can generate both types of torques. The speaker emphasizes the need for careful sample variation and measurement to understand the origins of these torques. The introduction of orbital effects in the context of spin-orbit torques is highlighted, with the potential for significantly enhancing the efficiency of torques, which is an exciting prospect for students and researchers in the field.

10:03

🌐 Orbital Hall Effect and Its Impact on Torques

The speaker explores the concept of the orbital Hall effect, which is analogous to the spin Hall effect but involves the generation of a transverse orbital current from a longitudinal charge current. This effect, along with the orbital Edelstein effect, can lead to the generation of orbital angular momentum flow, which is a promising area of research. The speaker also discusses the motivation for studying these effects, including the potential for large orbital Hall conductivities in certain materials and the possibility of avoiding the use of heavy, rare, and environmentally harmful materials in device fabrication.

15:04

📈 Measuring Orbital Effects Through Torques and Magnetoresistance

The speaker describes the experimental approaches to measuring orbital effects, focusing on torque measurements and magnetoresistance effects. The use of magnetic insulators to eliminate complications from charge currents within the material is highlighted. The speaker presents unpublished results showing a significant increase in torques when copper oxide is added to platinum, suggesting the generation of an orbital current that is converted into a spin current, which then acts on the magnetization. The talk also covers the measurement of spin Hall magnetoresistance effects and the potential for identifying orbital effects through these measurements.

20:05

🔍 Investigating the Mechanisms of Orbital Torques and Magnetoresistance

The speaker delves deeper into the investigation of orbital torques and magnetoresistance, discussing the use of different materials and the impact of their properties on the observed effects. The talk presents results from experiments with permalloy and copper, highlighting the differences in the length scales of the effects observed. The speaker also discusses the importance of understanding the physical origins of these effects, comparing them to standard spin Hall magnetoresistance and suggesting that the observed effects may be due to orbital Hall inverse or orbital Rashba Edelstein magnetoresistance.

25:06

🛠️ Enhancing MRAM Performance with Orbital Torques

The speaker discusses the application of orbital torques in improving the performance of Magnetic Random Access Memory (MRAM) devices. The talk presents results from experiments with ruthenium and niobium, which have large orbital Hall conductivities, and the use of platinum layers to enhance the conversion of orbital currents into spin currents. The speaker also addresses the challenges of anomalous torques and self-talks, and how they can be mitigated by using specific material combinations. The talk concludes with the significance of these findings for the development of MRAM devices and the potential for further research in this area.

🏆 Conclusions and Acknowledgments

In the concluding part of the talk, the speaker summarizes the key findings and contributions of the research, emphasizing the potential of orbital effects in enhancing the efficiency of torques and the importance of collaboration with theoreticians. The speaker acknowledges the work of the research team, including specific individuals who contributed significantly to the studies. The talk also credits external collaborators and funding sources that supported the research. The speaker invites the audience to a poster session for further details and expresses gratitude for the attention of the audience, looking forward to discussions.

Keywords

💡Spin-transfer torque

Spin-transfer torque is a phenomenon where the spin of electrons is transferred to the magnetization of a material, affecting its orientation. In the context of the video, it is a fundamental concept for understanding the limitations of current technology, as it is limited by an efficiency of one Bohr magneton per electron. The script discusses how this technology is already in the market but faces efficiency constraints.

💡Spin Hall effect

The Spin Hall effect is a quantum mechanical phenomenon where a charge current in a spin-orbit coupled system generates a transverse spin current. The video script uses this term to describe one of the mechanisms that can generate spin currents, which in turn can exert torques on the magnetization of a ferromagnet, contributing to the theme of exploring new methods for manipulating magnetization.

💡Orbital angular momentum

Orbital angular momentum refers to the angular momentum of a body arising from its motion in an orbit. In the video, it is highlighted as an alternative to spin angular momentum for transferring more than one Bohr magneton per electron, which is a significant opportunity for increasing the efficiency of torques in magnetic devices.

💡Damping-like torque

Damping-like torque is a type of spin-transfer torque that acts in the plane of the sample and is associated with the spin Hall effect. The script discusses how experimentalists can only measure the effective field of the torques, not distinguishing between damping-like and field-like torques directly, which is a key point in understanding the experimental challenges.

💡Field-like torque

Field-like torque is another type of spin-transfer torque that acts perpendicular to the plane of the sample. The video script mentions this term to illustrate the complexity of identifying the origin of the torques measured in experiments, as both damping-like and field-like torques can be generated by different mechanisms.

💡Orbital Hall effect

The Orbital Hall effect is an analog to the Spin Hall effect but involves the generation of an orbital current instead of a spin current. The script discusses this effect as a potential source of large orbital currents that can be converted into spin currents, which is a central theme in the exploration of new magnetic manipulation techniques.

💡Rashba-Edelstein effect

The Rashba-Edelstein effect is a phenomenon where a non-equilibrium spin density at an interface can lead to a torque on the magnetization due to the spin-orbit interaction. The video script refers to this as a mechanism that could potentially be harnessed for generating large torques, which is part of the broader narrative of seeking more efficient magnetic manipulation methods.

💡Magnetoresistance

Magnetoresistance is the change in electrical resistance of a material due to the influence of an external magnetic field. The script mentions this term in the context of measuring the effects of orbital currents, such as the Orbital Rashba-Edelstein magnetoresistance, which is a way to indirectly infer the presence of orbital effects in magnetic materials.

💡Magnetic insulator

A magnetic insulator is a material that exhibits magnetism but does not conduct electricity. The video script discusses the use of a magnetic insulator in experiments to isolate the effects of the charge current at the interface, which is crucial for understanding the mechanisms of torque generation without the complications of self-torques in metals.

💡Skermion

A skermion is a type of topological magnetic excitation. While the script mentions skermions, it does not focus on them, but the mention serves to highlight the breadth of research in the field of magnetism and the search for new phenomena that could be exploited in magnetic devices.

💡Niobium

Niobium is a chemical element used in the script as an example of a light, abundant, and relatively non-toxic material that could be used in experiments to study orbital effects. The script discusses the use of niobium to demonstrate the potential of light metals for generating large torques without the need for heavy or rare materials.

Highlights

The transition from theory to addressing experimental challenges in spin transfer torques.

The limitations of spin transfer torque efficiency and the fundamental limit of one Bohr magneton per electron.

Introduction of spin-orbit torques as a next step beyond spin transfer torques.

The role of spin Hall effect and the inverse spin galvanic effect in generating spin-orbit torques.

The difficulty in experimentally distinguishing between different types of torques due to measurement limitations.

The potential of orbital angular momentum in enhancing torque efficiency beyond the spin transfer torque limit.

Experimental results showing a 16-fold increase in torques with the addition of copper oxide.

The significance of the non-monotonic dependence of torques on the thickness of platinum.

The discovery of orbital Rashba-Edelstein magnetoresistance effects in certain material systems.

The importance of selecting appropriate ferromagnetic materials for efficient orbital to spin current conversion.

Demonstration of large orbital-induced torques in niobium and ruthenium systems.

The ability to manipulate iron cobalt boron using platinum layers to enable large torques.

The experimental validation of the absence of intrinsic large effects in nickel itself.

The theoretical and experimental collaboration that led to the discovery of enhanced orbital torques.

The practical implications for MRAM devices in utilizing iron cobalt boron instead of nickel.

The demonstration of orbital to spin conversion in permalloy itself without the need for additional layers.

The strong dependence on the ferromagnet in the orbital effects, highlighting a unique characteristic of orbital-induced phenomena.

Transcripts

okay yeah thanks a lot so it's great to

be here it's also great to see so many

faces live

in the last three years we had some

spice workshops where i saw 80 tiles of

zoom so this is a lot better seeing 80

people alive

in 3d that someone wrote not just 2d

yeah so so today i'm going to show some

experimental results and i'm kind of

going to kind of change gears a little

bit from a lot of really exciting theory

talks which you have seen to some of the

banal problems that experimentalists

actually face

um yeah so i'm local from mines and i'm

also have a co-affiliation at ntnu

so firstly also for the students why are

we doing all this complex stuff

i mean obviously it's exciting signs but

if you look at what is already in the

market you can see that actually there

are already been transfer talk devices

which you can buy

and they work by polarizing a current in

a polarizing layer and the current these

electrons which are spin polarized then

go to a free layer and change the

magnetization of the free layer and this

is really good it works you can buy it

it's actually a product out there but it

is also limited by an efficiency of one

h bar per electron for spin transfer

torque and that is of course something

that fundamentally

means that you need a certain amount of

current a certain amount of spins in

order to reverse the magnetization and

we heard from

the horn this morning if you go to low

magnetization materials this can

actually improve things but

fundamentally we are still limited for

every electron that has a spin of h bar

over two it can flip from plus h b over

two to minus h b over two so transfer

only one h bar

and the same also holds when you look at

things like domain bar motion or

skermion motion that when the electron

goes across a domain wall then in the

adiabatic limit it just switches its

spin from up to down so by one h bar so

you transfer one h bar onto the

magnetization of the domain wall and

that then limits the domain wall

velocity now of course we all know

because we hear that there is something

more than spin angular momentum and that

is orbital angular momentum

and so with this it was shown that you

can transfer more than one h bar per

electron if you have multiple

interactions often electrons say with

the domain wall as the electron passes

across it

so this is fundamentally a really

exciting opportunity that you can

increase and it's been shown

experimentally affected 10

in terms of the efficiency of the

torques compared to the previously used

spin transfer talks

so the spinomid torques which is kind of

the next step beyond spin transfer

torques this is something which you

already heard also again from a few

speakers before there are two

fundamental mechanisms which is the spin

hall effect there's a nice review from

hiro

and this is essentially

spin right electrons which are scattered

or with intrinsic uh moving to the top

interface and then they diffuse into a

ferromagnet and then they act on the

magnetization and now comes again the

experimental banality and that is what

we can measure is only the direction of

the exerted torques or even worse we

only measure effective fields which

either have the symmetry of a so-called

damping-like or field-like torque and

that's all we can provide you with so so

please theoreticians believe us we do

not measure

a spin hall effect torque or rush by

edelstein talk what we measure is an

effective field

and we can tell you the symmetry of the

field that's it and all the rest is

indirect interpretation

now in addition to this bulky effect

which is the spin hall effect you also

have this inverse spin galvanic effect

or rush by either shine effect where we

have a non-equilibrium spin density that

forms f4 current flowing at the

interface which then also exerts a

torque by exchange interaction on the

magnetization

so this type of torque that is exerted

can also have the two symmetries of

damping like oslanjeski like and now

there is a lot of literature and now for

the students some people call this

damping like talks lonzewski talker

anti-damping or in-plane or anti-dumping

which i thought was really nice um

i don't need to find out the paper i

said published paper where it's an

anti-dumping talk

um

okay and there's a field like which some

people also call the effective here the

rush ba

like perpendicular out of plane and so

on now what we are always interested in

is of course if we are fundamental

physicists to understand which talk

originates from which origin

and again we also have these orbital

effects that i'm going to talk about

next but again please remember that the

only thing we can provide you with is

the strength of the effective field

which is mini tesla current density so

melee tesla per amps per square meter

and the direction so the symmetry if

it's damping like or a field like torque

and now we need the input from the

theoreticians to actually see what do we

have to vary in our samples in order

then to understand where one talk

originates from and that is actually

quite tough when i started the business

with spin orbit torx in my group i don't

remember eight years ago or something

like that everyone was saying

spin hall effect generates a damping

like torque rushed by edelstrine effect

generates a feed like torque and that is

absolutely not correct so i think we

have meanwhile established that both

mechanisms can generate both torques

also theoretically but also that a lot

of simple experiments like varying a

thickness of a layer are not sufficient

because when you vary a layer you change

interfaces you change strain a lot of

things that can happen so it is a

difficult task but it's not to put you

off so for the students this is super

exciting because we suddenly have big

effects

i wasted my youth on permaloyd permaloy

has nice domain walls and you can move

them to spin transfer talk but it's a

very small effect the fields are

a few earths at 10 to the 12 m's per

square meter in spin orbit talks we

already have 10 times more we have milli

tesla per 10 to the 12 amps per square

meter and now with orbital talks i'm

going to show you another factor 10. so

we are 100 times better than where we

were at uh 20 years ago when i did my

phd

okay so that's kind of the introduction

now comes the interesting and difficult

signs

um you already heard all of this so in

an analogy to the spin-hole effect that

generates a transfer spin current from a

longitudinal charge current we also have

an orbital hall effect that generates a

transverse orbital current from a loot

longitudinal charge current and we have

an analogy to the inverse being galvanic

effect or the spin rust by edelstein

effect that generates a non-equilibrium

spin density at the interface we can

generate by the orbital rushed by either

side effect and non-equilibrium orbital

density at the interface

so okay

now again remember we measure two

torques and now we already have four

origins so actually there is a lot of

scope for you know getting things wrong

um but of course before we actually go

into the details to see what we can

learn let me just uh

you know my summary of what we heard

from hyun woo

yesterday so

in the 3d metals typically we have in a

ground state orbital quenching but when

we actually have an electric field or

also some other excitation we can get

orbital angular momentum flow or

accumulation and that makes it really

exciting so even though

in most typical 3d metals and apparently

what i tell my students in my magnetism

lecture course is not wrong in the

ground state the orbital moment is

largely quenched when you excite it you

can generate a strong orbital uh angular

momentum flow so i should also say that

i'm absolutely new to this field so

there's already tens of years of

experience in this there's some original

work here of course don't work work from

a couple of years ago so we came to this

experimentally via a lot of motivation

from dongbok and jerome kruzov to

actually understand some of our results

so that's also really nice to be

motivated by theoreticians to look for a

new effect

now one very motivating slide which i

took from drunk work is this one here

where we compare the conductivity the

spin whole conductivity in the orbital

hole conductivity and then we saw that

if you compare theoretical calculation

of platinum and manganese we can get

like an order of magnitude more orbital

hole conductivity that obviously is

extremely motivating because i can tell

you after a lot of frustration in

permaloya very small effects measuring

large effects is simply better fun also

for the students

so

yeah that motivated us um and there's

this this work here where they

calculated the orbital conductivity and

found that for some materials which you

know for us were not very exciting

people for

very very large or potentially large

orbital conductivities

now in addition to the bulk orbital hall

effect we have also the orbital rush by

edelstein effect where you have a chiral

double angular momentum texture in k

space that induces the orbital angular

momentum this an accumulation of orbital

angular momentum which is analogous to

the spin rush beta stein effect and what

is exciting is that it can exist in

light metals and acidic structures like

copper copper peroxide without the need

for strong spinal coupling

and one has to also point out again also

for the students that

if you talk to companies like intel

they're not super excited about using

tons of iridium ruthenium or platinum

because it's rare it's heavy

it's also

sometimes toxic to the environment so

cadmium is also really nice material for

some experiments but i can tell you no

no if you talk to the european

commission they're not very happy to

have cadmium in any of the devices that

you develop i think if you run write a

grant proposal we're going to do mercury

and cadmium

might have some issues but

so that is another motivation that i

think we should not underestimate that

there is not necessarily now a need for

heavy environmentally detrimental and

rare and also expensive atom species so

i'll show you later also some results

that we have on niobium which is a

relatively abundant and relatively light

material

again there's some original work

here more than 10 years ago

and if you're interested have a look at

that

now

this is all great i mean this you

probably all heard about it just with a

little spin so now comes the

experimental problem how can we identify

what can we measure

and there are three things and we

already heard about a really nice

measurement which is the curve effect

measurement where you just have the

single material i think in this case

titanium i think this is a really nice

approach of course as we saw it's also

tough i mean nanowrite i mean maybe

theoreticians don't realize but

measuring nanowright rotation is not a

lot so it's not so easy

um so it's really a nice feat um but we

are actually interested in

measuring the uh orbital effects by

measuring torques and measuring magneto

resistance

um and i think these are the two things

that i'd like to walk you through and

also show you some unpublished hot

results

that we just got in the last couple of

months

so let's start with the torques so the

first

thing that i would like to show is that

there are of course ways to quantify

torques that we use for spin orbit

torques and then there are ways to

indirectly infer where they come from

so if you look at

these two pictures here

then in this case here we have a

light metal let's assume it's niobium or

whatever where we generate by a

longitudinal charge current a transverse

orbital current and this orbital current

then enters a ferromagnet and now you

know what should it do there are two

ways that you can actually act it can

directly interact with the orbit of the

ferromagnet or as we heard also

yesterday we can convert the orbital car

into a spin current and the spin current

then acts on the magnetization

so in the end what we're measuring is a

change is an effective field acting on

the magnetization you can also have the

or the opposite which is orbital pumping

so this i think was first shown in this

paper here again i think we should give

a lot of credit to the people that

actually did this a little bit

underestimated i think for four or five

years no one gave a about that pair

very few people cared about that paper

until people actually understood how

groundbreaking this was and then you

know as you can see we were like four

years later

um

so the experiment that we wanted to do

is this one here we want to measure the

torques acting on a magnet

and the first experiment we did was to

work on a magnetic insulator

now if you have a magnetic metal you

have a lot of things happening you have

current flowing in the metal you get

self torques anomalous torques

so i i really like magnetic insulators

not because they're easy to make they're

actually not easy to make but because

you don't have the problem that you have

any charge kind of flowing in them so

anything that happened has to happen by

the charge current at the interface

there's no electrons entering into the

magnetic material so what we chose is

sodium ion garnet platinum of course it

also has germion so you know it's an

amazing material and skeletons are very

hot and everyone loves skermions but i'm

not going to talk about it but just

partly we had this material at hand

because we were studying in this paper

here actually skermions in this thalium

iron garnet

then we put platinum on top

to start us studies without torques and

yes so this is a measurement of the

spinova talks and there are different

ways to measure that if you're

interested here's the paper but what is

important is if we vary the thickness of

a platinum

we see that it increases and levels off

and that's very typical for all spin

orbit torque measurements

so you have initially an increase up to

the thickness of the spin diffusion

length in the platinum and then it

levels off because essentially all the

spin current that is generated in the

platinum then

a further away from the interface

doesn't get to the interface and you

maximize the efficiency of the torque

here

so this is actually pretty much what we

get in yik platinum thorium iron garnet

platinum lots of garnets platinum with

something like a spin diffusion length

of 1.8 nanometers in the platinum that's

all standard

now my very smart student chile ding he

put copper copper oxide on top partly

motivated by this existing material

and i was very skeptical because firstly

copper is you know light secondly copper

oxide is insulating so somehow you know

i don't see why it should do anything

useful

and thirdly because you know he just let

it rot at air which i thought was not a

very well defined process but then this

is what happened

he saw that if you put copper oxide on

top you get a 16 fold increase of the

acting torques so sometimes for the

students do something crazy and you know

even if you're a supervisor or don't

tell your supervisor i only tell him

when you get the curve

um so uh i think that this you know he

came back with that curve like that you

know something is wrong so we did a lot

of tests we checked you know what is the

conductivity of the copper oxide we

removed the platinum see what happened

but actually it's robust

and we saw that for

a certain thickness of the platinum you

actually have a strong increase in the

torques and you know not a strong but a

16-fold increase that's gigantic i mean

you know we were always fighting for a

factor two here just you get for free

effective 16.

so what is happening

interpretation was that

in the copper copper peroxide you

generate an orbital current at the

interface and now this orbital current

in the platinum is actually converted

into a spin current and the spin current

then generates a spin-orbit torque that

acts on the thorium iron garnet

and what is interesting is of course if

you look at here there's non-monotonic

dependence and that can be described

quite nicely because of course when you

have very thick platinum then eventually

the orbital current that is converted in

the first one point uh something

nanometers of the platinum

then simply doesn't get any more spin

current to the fullium iron garnet so if

you make a lot of platinum

of course nothing will happen if you put

one meter of platinum and you put some

copper oxide on top nothing will exit at

the bottom interface but for thin

platinum layers you have an extremely

efficient torque generation and that

means you must have a huge orbital

current because you now have two times

things that need to happen you generate

the orbital current it needs to be

converted to a spin current and the spin

current then acts on the magnetization

so fundamentally one would say should be

less efficient just means there's a huge

orbital current that is being generated

yeah and then we could fit this and get

the length scales of the orbital to char

orbital to spin current conversion in

about one nanometer

now what is also nice about these iron

garnets is they have typically nice

spin-hole magnet resistance effects and

then we measure the spin hole magnitude

resistance in the system and we also

found this peak at something like one

point something nanometers of platinum

thickness so here again this is the

series a is a series without the copper

copper oxide

and then we see there is a typical

increase and then decrease just because

you get shunting

and

if you put the copper sorry in the

series b the copper copper oxide on top

you have this huge increase the 16 fold

16 fold increase and then the reduction

as you get more shunting and also we put

some actually all the credit to chile he

did it

put also mgo to protect the copper from

oxidation so not let it rot in air and

then you see there is no increase at all

so it's really the copper copper oxide

that makes a difference not the copper

in this case

and we should say that you know a lot of

the explanation was supported by donbook

and euron

okay so this is the first thing we can

do so we can generate huge torques by

putting some really crazy stuff on top

of platinum

the second thing we did is we wanted to

look at magnetoresistance effects to see

if we can actually also obtain magnesium

resistance effects in these type of

systems and here we remove the platinum

so what we did is we just did permalink

copper you know this is from my youth is

like permaloy and so i said how about we

just put parapromelo and then copper

copper oxide and see if we actually need

this conversion layer

and then for those of you that are into

magnetic resistance effects there are

three scans that we can do we can rotate

the field in plane we can rotate the

field out of plane perpendicular to the

current or out of plane in the current

direction and then if we do this beta

scan then we get the smr type spin hall

magnetoresistance type of resistance in

this case is orbital rashba edenstein

magneto resistance and we do this now

for different thicknesses but this time

we vary the copper thickness so the

permeable is fixed but we vary the

copper thickness and what we see is that

actually we get an increase with

increasing copper

and again we have nothing on top so we

get copper copper oxide

and then as we increase the copper

thickness eventually the copper oxide

doesn't oxidize through and we get

shunting and the effect goes down again

now what is interesting about this is

if we now compare

the different thickness of permaloy and

the different thickness of the copper

we can see that if we just put permeable

platinum we again get an increase of the

spin hall magneto resistance for the

first two nanometers or so and then a

decrease

but if we put copper oxide on top we get

a different length scale we get a peak

at something like seven nanometers

compared to three nanometers here and

the length scale of 5 nanometers

compared to 1.83 nanometers here and

that shows the actual magneto-resistance

effect must have a different mechanism

must have a different physical origin

so

here it's standard um inverse spin hall

effect so spin hole magnitude resistance

and here we have orbital hall inverse or

orbital rush by edenstein

and inverse orbital rochelstein effect

so this i think is really exciting

because now we have two means to

determine

the possible orbital effects that can

occur one is measuring torques one is

measuring many resistance effects

now

copper copper oxide is great

effects are big but it's also badly

defined you know theoreticians don't

like it you know if you talk to don't

work he doesn't like calculating copper

oxide because we cannot tell him what we

really have you know we don't have good

copper copper oxide what we have is

copper that rots in air we also do it in

plasma but then anyway so

you know then we have peter peter comes

up with all these amazing calculations

so this is from this archive paper here

and then you know he compares and it's

not only him obviously there's this work

here there's don rook's work

and and many others and then he

calculates the spin hole conductivity

and the orbital hall conductivity and

then you just go through this and you

just pick something where you see a peak

like athenium and you select let's do

that ruthenium is great we have

ruthenium in the chamber and therefore

we can just deposit it and you know we

don't need this oxide which is badly

defined it's also a little bit

frustrating if a paper like this comes

out because this is like you know 10

years of work for an experimentalist to

go through this

and maybe three months of calculation

anyway so what we did is we picked

ruthenium and niobium because they have

large orbital hole conductivity but they

have very small spin hole conductivity

so that's what we did and also don't

work did the calculation uh in a bit

more detail for niobium ruthenium and we

see huge

orbital hall conductivity and in red

very little spin hall conductivity and

then we did measurements and now all the

credit to arnhem ana bose he's a postdoc

working with me and the paper that we

just finished is from him and he will

show details tomorrow in uh the poster

so what we measured was niobium

nic iron cobalt boron and a niobium

nickel and we measured the acting

torques

and we first measured with iron cobalt

boron which is our standard system for

tmr junctions and the torque is super

small we see virtually no torque

and then we measure on nickel and we

have a huge increase in the torque and

now this is something important again

for the students this does not happen

for spinova torques for spin orbit

torques the effect is big on all

ferromagnets or small on all

ferromagnets here there is a huge

difference in that we have a major

difference whether we act on a different

ferromagnet or nickel or iron cobalt

boron

so the same holds for ruthenium

again withinium and iron cobalt boron

very little effect and ruthenium nickel

huge effect and the interpretation is

that you need to convert the orbital

current into a spin current and that

works well in nickel

but it does not work well in iron cobalt

boron so i think here is a huge field

also for students to invest yourself

experimentally trying out different

ferromagnets to see what ferromagnet

makes the best conversion from an

orbital current to a spin current by the

way this was measured by sdf homogeneous

in case someone's interested ask arnob

about details tomorrow

and now i think we had yesterday the

question from kung jin li how about

anomalous talks and and um

self-talks so what we did is we actually

then

sandwiched the nickel between ruthiem on

both sides so that we have equal

interfaces and then the effect goes away

so it goes away from niobium and

ruthenium so there is no large intrinsic

effect in the nickel itself so this is i

think also a very important check now i

said okay

there's no

strong talk on iron cobalt boron but

there's a strong talk on nickel can we

make actually iron cobalt boron to be

manipulated

and well it turns out we can because

what we can do is we can again inject

our platinum layer

so now we have a more complex system

where we have ruthenium not directly on

iron cobalt boron but with a platinum

layer and we again put a second platinum

layer on the other side to cancel out

the spin hole conductivity and if you

just do platinum iron cobalt boron

platinum then the effect is virtually

zero but now we put ruthenium on one

side and we're going to get a large

torque

so now we are able to generate a large

orbital carbon in ruthenium

convert it to a large spin current and

platinum and add spin current works

effectively on the iron cold boron and

that is important because companies

don't like nickel nickel is not a very

nice material for an amram device they

want cobaltine boron because they want a

good mgo tunnel junction afterwards and

so the fact that we can use just one

nanometer of platinum to get again this

very large torque allowing us to act on

the iron cobalt boron or coal dye and

boron is super important because then

this allows to actually switch iron

cobalt boron which is what we want to do

for an mram device

now all the credit to anup and his

poster tomorrow so please do see his

poster and now i see my time's already

up so this is the people that did all

the work i just get to talk

and travel around the world or travel

not so far around the world and give

talks so these are the people that were

involved in particular arnab fabian and

ross saoin yen chile ding sven becker

kyung jin lee

klaus rabb and lucia and the others as

well in particular three staff

scientists martin jordan gerhard jacob

and our visiting professor hartmut zabal

this is what we looked like last year

and oh sorry

so this is what we looked like last year

now this is what we look like this year

things are going uphill also a lot of

credit go to our colleagues from picking

university who worked with us

together with chile on these first two

prls and also a lot of credit go to the

theoreticians that motivated us to get

into that direction so don vuk and frank

from eulek and yura also from euleg and

mainz and then colleagues in sendai

sydney and also the group of hiro helen

gomonae karen averso city has now moved

to duisburg and people wear measurements

as well as some of the funding and some

work on the iron garnets with caroline

ross and yeah most of the work was

funded by the germ research foundation

as well as the eu and also some from

companies

and uh yeah with that i summarize so i

mean this is the banal

thing analogous to the spin hall effect

in the rush beta shiny effect are the

orbital versions but they are not easy

to identify as an experimentalist it's

not straightforward to say what we

measure is due to one or the other

however there are ways to indirectly

infer it by looking at multiple samples

with different stacks and understanding

how orbital currents can be converted

into spin currents in heavy metal layer

or layers that have large spin-over

coupling and this was shown in this

first pll here

we find this very very strongly enhanced

orbital talks which i found extremely

surprising and then exciting because it

means we can go beyond what is the state

of the art

we measured the orbital rush by eighteen

strike magnetoresistance where we saw

that we can even get away from the

platinum and use a stack just comprising

a ferromagnet and a light metal layer so

we have orbital to spin conversion in

the permaline itself so not necessary to

have another layer

however if you do want to have iron

cobalt boron switch you need this

additional layer and there we find that

in well-defined niubium ruthenium where

we know the crystallite size we know the

crystalline structure we can tell don't

work what we actually have so you can

calculate something reasonable uh we get

a strong dependence on the ferromagnet

which is unique to at least as far as i

can see to this orbital effect is not

present in spin-orbit effects

and yeah please see on us poster

tomorrow and there's a nice review that

don book wrote and i contributed a

little bit to it and with that i thank

you for your attention and i'm looking

forward to the discussions